Botanical Sciences 99(3): 542-559. 2021 Received: November 15, 2020, Accepted: February 2, 2021 DOI: 10.17129/botsci.2795Genetic structure of threeOn line Mexican first: May 26, 2021 species

Genetics / Genética

A population genetics study of three native Mexican woody bamboo species of (: Bambusoideae: Bambuseae: Guaduinae) using nuclear microsatellite markers Estudio de genética de poblaciones de tres especies de bambúes leñosos nativos de Mexico del género Guadua (Poaceae: Bambusoideae: Bambuseae: Guaduinae) utilizando microsatélites nucleares

Jessica Pérez-Alquicira1,2, Stephanie Aguilera-López1, Yessica Rico3,2, Eduardo Ruiz-Sanchez*1,4

1 Universidad de Guadalajara, Laboratorio Nacional de Identificación y Caracterización Vegetal (LaniVeg), Depto. Botánica y Zoología, Centro Universitario de Ciencias Biológicas y Agropecuarias, Universidad de Guadalajara, Zapopan, Jalisco, México. 2 CONACYT, Ciudad de México, México. 3 Red de Diversidad Biológica del Occidente Mexicano, Instituto de Ecología, A.C., Pátzcuaro, Michoacán, México. 4 Universidad de Guadalajara, Departamento de Botánica y Zoología, Centro Universitario de Ciencias Biológicas y Agropecuarias, Zapopan, Jalisco, Mexico. *Author for correspondence: [email protected]

Abstract Background: Sporadic flowering contributes significantly to genetic diversity and connectivity among populations. Woody present sporadic or gregarious flowering patterns with long flowering cycles. In this study, we analyze the genetic diversity of three Guadua species distributed along the Gulf of Mexico slope that have different patterns of flowering. Questions: (1) Are the three Guadua species genetically differentiated? (2) Does the vulnerable species G. inermis have low levels of genetic diversity? (3) What is the relative contribution of geographic and environmental factors to the genetic structure of G. inermis? Species studied: Guadua amplexifolia, G. inermis and G. tuxtlensis. Study site and dates: During 2014 and 2015, we collected samples of G. inermis in Puebla and southeastern Mexico, G. amplexifolia in Vera- cruz and Oaxaca, and G. tuxtlensis in southern Veracruz. Methods: We successfully amplified five of nine SSR markers, and genotyped a total of 155 samples. Results: The three Guadua species were genetically differentiated. For G. inermis, we found high levels of population genetic diversity, which are relatively higher than those of other monocot species. Genetic differentiation was high and three groups were detected: north, central and south. We found a significant association between genetic distances and the maximum temperature of the warmest month, but not with geo- graphic distance. Conclusions: Our study is the first to analyze levels of genetic diversity in Mexican bamboos and confirms their taxonomic identity.G. inermis has a strong genetic structure, even when populations are geographically close. Keywords: Bamboos, genetic diversity, genetic structure, sporadic and massive flowering, polyploid. Resumen Antecedentes: La floración esporádica puede influenciar la estructura genética de las poblaciones. Los bambúes presentan patrones de floración esporádica o gregaria. En este estudio analizamos la diversidad genética de tres especies de Guadua con diferentes patrones de floración distri- buidos en la Vertiente del Golfo de México. Preguntas: 1) ¿Están las tres especies de Guadua (G. inermis, G. amplexifolia y G. tuxtlensis) genéticamente diferenciadas? 2) ¿Presenta G. inermis bajos niveles de diversidad genética? 3) ¿Cuál es la contribución relativa de los factores geográficos y ambientales en la estructura genética de G. inermis? Especies de estudio: Guadua amplexifolia, G. inermis y G. tuxtlensis. Sitio y años de estudio: En 2014 y 2015, colectamos muestras de G. inermis en Puebla y sureste de México; G. amplexifolia en Veracruz y Oaxaca, y G. tuxtlensis en el sureste de Veracruz. Métodos: Amplificamos exitosamente cinco marcadores SSR de un total de nueve y genotipificamos 155 muestras. Resultados: Las tres especies analizadas están genéticamente diferenciadas. G. inermis presentó altos niveles de diversidad genética, y superio­ res a otras especies de monocotiledóneas. Los niveles de diferenciación genética fueron altos y se detectaron tres grupos: norte, sur y centro. Encontramos una asociación significativa entre las distancias genéticas con la temperatura máxima del mes más caliente, pero no con la distancia geográfica. Conclusiones: Nuestro trabajo es el primer estudio que evalúa los niveles de diversidad genética en bambúes mexicanos y confirma su identidad taxonómica. G. inermis tiene una estructura genética alta, incluso cuando las poblaciones están cerca geográficamente. Palabras clave: Bambúes, diversidad genética, estructura genética, floración esporádica y masiva, poliploidía.

This is an open access article distributed under the terms of the Creative Commons Attribution License CCBY-NC (4.0) international. https://creativecommons.org/licenses/by-nc/4.0/

542 Pérez-Alquicira et al. / Botanical Sciences 99(3): 542-559. 2021

Understanding the processes leading to speciation and tween populations, leading to diffuse flower temporality biodiversity in is one of the main research topics in within a species; 4) Sporadic flowering, random or other evolutionary biology. Flowering time has a strong effect non-gregarious patterns of flowering (Troup 1921, Janzen on gene flow among populations (Fitter & Fitter 2002). 1976, Gadgil & Prasad 1984, Banik 1998, Franklin 2004, For example, floral synchrony results in the simultaneous Bhattacharya et al. 2009, Zheng et al. 2020). Understand- and massive flowering of individuals, offering abundant ing how asynchronous flowering phenology can influence floral resources that can attract numerous pollinators and patterns of gene flow in woody bamboos is crucial to elu- favoring outcrossing rates (Domínguez & Dirzo 1995, cidating its role in population genetic divergence and spe- Rodríguez-Pérez & Traveset 2016). In the same way, ciation. there is evidence that flowering synchrony increases re- The genus Guadua Kunth is one of the six genera of productive success in wind-pollinated species (Bogdzie- the subtribe Guaduinae (Judziewicz et al. 1999, Clark et wicz et al. 2020). In contrast, asynchrony or sporadic al. 2015, Tyrrell et al. 2018). It includes 33 species and is flowering might reduce the availability of mates and thus one of the most diverse genera in the subtribe (Clark et reproductive success in both insect and wind-pollinated al. 2015, Ruiz-Sanchez et al. 2021). Species of Guadua species (Domínguez & Dirzo 1995, Bogdziewicz et al. are distributed from Mexico to Argentina, most of the spe- 2020). Asynchrony in flowering time across populations cies occur in South America, and this region is consid- should result in lower levels of gene flow and greater ge- ered its center of diversification (Medina & Medina 1965, netic structure, even for populations that are geographi- Judziewicz et al. 1999, Londoño 2001, 2011, Clark et al. cally close together (Kirkpatrick 2000, Reisch & Posclod 2015). Mexico has seven species of Guadua: G. aculeata 2009). Rupr. ex Fourn., G. amplexifolia J. Presl., G. longifolia Geographic distance and the environment represent (E. Fourn.) R.W. Pohl, G. paniculata Munro, G. inermis key components influencing species’ genetic structure Rupr. ex Fourn, G. velutina Londoño & L.G. Clark, and G. (Rousset 1997, Wang & Bradburd 2014). Several species tuxtlensis Londoño & Ruiz-Sanchez. The last three spe- exhibit a significant pattern of isolation by distance (IBD), cies are endemic to Mexico, and except for G. paniculata, which suggests that genetic differentiation increases with which is distributed along the Pacific coast of Mexico, geographic distance. On the other hand, isolation by envi- the other species are distributed along the Gulf of Mexi- ronment (IBE) has been proposed to describe the relation- co in the states of Campeche, Chiapas, Hidalgo, Oaxaca, ship between environmental heterogeneity and the spatial Puebla, San Luis Potosí, Tabasco, Tamaulipas, and Vera- distribution of genetic variation (Wang & Bradburd 2014). cruz (Londoño & Ruiz-Sanchez 2014, Ruiz-Sanchez et Thus, genetic differentiation among populations increases al. 2020). Most of the species are used to build houses, with their environmental differentiation (Wang & Brad- fences, kiosks, and to make handicrafts (Londoño & Ruiz- burd 2014). In contrast, limited dispersal constrains gene Sanchez 2014). flow among populations and reduces the possibility of Few studies have reported the flowering cycles ofGua - counteracting genetic drift (Slatkin 1993, Rousset 1997). dua species. Guerreiro (2014), Guerreiro et al. (2020) Woody bamboos (Poaceae: Bambusoideae: Bambuse- and Vega & Hernández (2008) recorded massive flower- ae) are a large taxonomic group with notable economic ing events and estimated flowering cycles for three South and ecological importance. Flowering in woody bamboos American Guadua species: Guadua chacoensis (Rojas) is peculiar because many of them remain in a vegeta- Londoño & P. M. Peterson (28-31 years), G. paraguayana tive phase for a long period of time, as long as decades Döll (38 years), and G. trinii (Nees) Nees ex Rupr. (30-33 or a century, followed by mass synchronous flowering years). Liebsch & Reginato (2009) estimated cycles of 27- and subsequent death (Janzen 1976). Other woody bam- 28 years between flowering events forG. sarcocarpa Lon- boos are characterized by sporadic flowering with only doño & P. M Peterson and G. weberbaueri Pilg, while spo- a few individuals reproductively active in their popula- radic flowering and massive events have been recorded in tions. Flower phenology in woody bamboos falls into G. angustifolia Kunth (Londoño 2002) and only sporadic four spatio-temporal patterns: 1) Flowering distribution, flowering in G. inermis Rupr. ex Fourn (Aguilera López with a small percentage of individuals flowering the first 2020). or second year before and after the main flowering; 2) Herbarium records and field collections show that G. Flowering wave, with gregarious flowering occurring in amplexifolia has massive flowering cycles of unknown pe- patches in successive years; 3) Variation in periodicity be- riodicity. We collected flowering plants of G. aculeata in

543 Genetic structure of three Mexican bamboo species

2012, and some from Puebla state that were flowering in G. amplexifolia occurs from Mexico to Colombia. Culms 2020 (Morocho pers. comm.). G. inermis with flowers have 10-15 m tall, 6-10 cm in diameter, solid at the base, and been collected in different localities of Veracruz in differ- hollow in its distal portion. Flowering phenology is mas- ent years (2009, 2012, 2014, and 2015), and from isolated sive. The distribution of G. tuxtlensis is restricted to Los blooming individuals (Ruiz-Sanchez pers. obs.). Finally, Tuxtlas, Veracruz in Mexico. The culms are 10 to 20 m flowering has not been recorded in G. tuxtlensis (Londoño tall and 8 to 14 cm in diameter. This species has hollow & Ruiz-Sanchez 2014). culms, walls are thick, thorny branches from the first node, Even though some bamboo species are of economic culm leaves are persistent basally (Figure 1). Flowering and ecological importance, there is little information on has not been recorded for this species. According to Guo their patterns of genetic diversity and structure in Mexico. et al. (2019), Neotropical woody bamboos are tetraploid Knowledge of the genetic diversity resources of woody and wind-pollinated species. bamboos would provide valuable information for conser- vation strategies as the habitat of bamboos is highly vul- Species sampling. We collected fresh leaves from these nerable due to human practices such as livestock farming. three Guadua species during 2014 and 2015. For G. iner- In particular, the size of G. inermis’ populations are small mis, we collected from thirteen populations in Veracruz, in comparison with those of other Guadua species (i.e., Chiapas, and Puebla. The Puebla population is probably G. velutina) and, according to Ruiz-Sanchez et al. (2018), a cultivated population, because the native distribution of the projected distribution of G. inermis in 2050 could be G. inermis does not include this state. G. amplexifolia was reduced by as much as 42 % as a consequence of global collected from three populations in Veracruz and Oaxaca, warming. Increasing habitat fragmentation, habitat vul- and G. tuxtlensis from two populations in southern Vera- nerability due to climate change, and the sporadic flower- cruz (Appendix 1, Figure 2). In total 155 individuals were ing of G. inermis will threaten the prevalence of this spe- analyzed, and the number of individuals collected per site cies. Therefore, analyzing its genetic diversity is crucial to varied from 2 to 16 (Appendix 2). Foliar tissue of adult develop effective conservation efforts. and juvenile plants was collected from plants separated In this study, we used nuclear microsatellites for three by at least 10 m to reduce the possibility of sampling the Guadua species (G. amplexifolia, G. inermis, and G. tux- same genotype. The material was preserved in silica gel tlensis) to answer the following questions: (1) Are the prior to DNA extraction. GPS coordinates were recorded three Guadua species genetically differentiated? (2) Does for each population. the vulnerable species G. inermis have low levels of ge- netic diversity? (3) What are the relative contributions of DNA extraction and microsatellite genotyping. Genomic geographic and environmental factors to the genetic struc- DNA was isolated from 100 mg of leaf tissue following ture of G. inermis? Doyle & Doyle’s (1987) CTAB procedure. The DNA was dissolved in 100 µL of Milli-Q water. We tested seven Materials and methods microsatellite primers developed for Guadua angustifolia (Pérez-Galindo et al. 2009) and two developed for Aulo- Study species. Guadua amplexifolia, G. inermis, and G. nemia aristulata (Abreu et al. 2011). PCR reactions con- tuxtlensis are morphologically similar, and some of their tained 2.5 μl of a multiplex solution, 5-100 ng of DNA populations are sympatric within the state of Veracruz template, and 0.3 μM of each primer. The final volume was (Londoño & Ruiz-Sanchez 2014, Ruiz-Sanchez et al. 5.5 μl. PCR reactions were performed using an AERISTM 2015, 2020). The first two species are the most morpho- thermal cycler (Esco Healthcare, Singapore) under the logically similar and are difficult to distinguish. It has been following conditions: 96 ºC - 15 min; 35 cycles of 94 ºC suggested that G. tuxtlensis is of hybrid origin and one of - 30s, a gradient of annealing temperature from 50 to 60 ºC the potential parent species could be G. inermis (Londoño - 1:30 min, 72 ºC - 1 min. The final extension lasted 30 & Ruiz-Sanchez 2014) (Figure. 1). Guadua inermis is min at 60 ºC. We successfully amplified five primers. The endemic to Mexico, occurring in the states of Campeche, forward primers FJ444929 and FJ444936 were labeled Chiapas, Oaxaca, Tabasco, and Veracruz; in tropical sub- with HEXTM, and the forward primers Aar12, FJ476075, deciduous forests (Cortés-Rodríguez 2000, Londoño & and FJ444930 were labeled with 6-FAMTM dyes. The Ruiz-Sanchez 2014). This species has culms 4 to 12 m tall, PCR amplification for labeled primers consisted of 2.5 μl 3 to 10 cm in diameter, and solid or thick-walled culms. of a multiplex solution, 5 - 100 ng of template DNA, and

544 Pérez-Alquicira et al. / Botanical Sciences 99(3): 542-559. 2021

Figure 1. Morphological differentiation in three Guadua species from eastern Mexico. A-B. Guadua tuxtlensis, C-D. G. amplexifolia, and E-F. G. iner- mis. Photos by Eduardo Ruiz-Sanchez.

545 Genetic structure of three Mexican bamboo species

0.5 μl of the primer mix. The primer mix included 2 μM Genetic diversity analyses. We estimated the average for each primer and two pairs of primers per reaction were null allele for each locus based on De Silva’s method combined except for one of them (FJ444929 + Aar12; (De Silva et al. 2005) implemented in Polysat (Clark & FJ444936 + FJ444930; and FJ476075). PCR temperature Jasieniuk 2011). We calculated genetic diversity statis- o cycling conditions were as follows: 96 C for 15 min; 35 tics, including observed (HO), Nei’s gene diversity cor- o o o cycles of 94 C - 30 s, 57 C - 1:30 min, 72 C - 1 min. The rected for sample size (HE), number of alleles per locus o final extension included 60 C for 30 min. We diluted each (NA), and effective number of alleles per locus (NAe) PCR reaction (1:30) and sent them to Illinois University, using the software Spatial Pattern Analysis of Genetic UIUC Core Sequencing Facility to conduct the fragment Diversity (SPAGeDi) v1.3a (Hardy & Vekemans 2002). analysis using GeneScanTM-500 LIZ Size Standard. Gen- We also estimated allelic richness using the rarefaction eMapper v4.1 (Applied Biosystems) was used to carry out method implemented in the Allelic Diversity Analyzer genotype scoring. Because G. inermis is a tetraploid spe- (ADZE) (Szpiech et al. 2008). This method trims un- cies, the genotype configuration for each allele was de- equal sample sizes to the same standardized sample size termined using the MAC-PR method, microsatellite DNA (g), taking a value that is less than or equal to the small- allele counting - peak ratios (Esselink et al. 2004). This est sample size for the whole dataset. We set g = 4 to method infers the number of alleles as a function of the have a more comparable measurement of genetic diver- peak area for each individual and each locus. sity among populations.

Figure 2. Collecting sites of three Guadua species in eastern Mexico.

546 Pérez-Alquicira et al. / Botanical Sciences 99(3): 542-559. 2021

Genetic structure analyses. Genetic differentiation among values (FST/1 - FST) were used to obtain the pairwise ge- the populations of all three species was estimated using the netic matrix. The geographic distance matrix was obtained

FST statistic implemented in the software SPAGeDi v1.3a using the great-circle distance in the geosphere R package

(Hardy & Vekemans 2002). In addition, the GST statistic v1.5.10 (Hijmans 2019). We obtained the 19 environmen- was calculated for each species using Genodive v 2.0b27 tal variables from WorldClim (www.worldclim.org/) with (Meirmans 2020). We performed assignment genetic anal- a resolution of 30 arcsec (1 km2). We performed a PCA on ysis using a Bayesian method implemented in the software all bioclimate variables. The variables most strongly cor- STRUCTURE v2.3.4 (Pritchard et al. 2000). We included related with the first three PCA axes containing the highest an admixture model and uncorrelated allele frequencies variance were selected. We checked for multicollinearity because we analyzed three bamboos species. Additionally, among the selected variables with the variance inflation we carried out a second analysis including only the popu- factor (VIF) and excluded variables with VIF > 10 using lations of G. inermis using the admixture model and cor- the vifstep function in the usdm R package v1.1.18 (Naimi related allele frequencies. We ran the analysis with 1 × 106 2015). The bioclimate variables selected were isother- Markov chain Monte Carlo (MCMC) steps and a burn-in mality (Bio 3), max temperature of warmest month (Bio of 1 × 105 steps, with 10 replicates for each K (number of 5), mean temperature of warmest quarter (Bio 10), pre- clusters), where K = 1 to 10. To determine the most appro- cipitation of wettest month (Bio 13), and precipitation of priate value of K, the average and standard deviation (SD) coldest quarter (Bio19). First, we tested for IBD using the of the likelihood of each model were used to calculated Mantel test with 1000 permutations in the vegan package ∆K (Evanno et al. 2005) using Structure Harvester (Earl (Oksanen et al. 2013). Then the statistical relationships of & vonHoldt 2012). The visual output of the STRUCTURE IBD and IBE with the genetic distance matrix were evalu- results was generated using DISTRUCT (Rosenberg 2003) ated with maximum‐likelihood population effects model and CLUMPAK (Kopelman et al. 2015). (MLPE; github.com/nspope/ corMLPE; Clarke et al. We also ran a Principal Components Analysis (PCA) 2002). This approach explicitly accounts for the non-inde- in the R package adegenet (Jombart et al. 2010). PCA is a pendence of values in regressions run on distance matrices multivariate method free of Hardy-Weinberg and linkage (Clarke et al. 2002). Parameter estimation was performed disequilibrium assumptions. We carried out an Analysis of using restricted maximum likelihood (REML). To select Molecular Variance (AMOVA) by grouping populations among competing univariate models, we utilized infor- by species and only for G. inermis populations without mation theoretical criteria such as the Akaike information a specific hierarchical structure. These analyses were run criterion corrected for sample size (AICc) and the Akaike in the popr R package with significant values determined model weights (wi) calculated using the dredge function after 20,000 replicates (Kamvar et al. 2014). from the MuMIn v1.4 package (Bartoń 2018).

Isolation by Distance (IBD) and Isolation by Environ- Results ment (IBE) in G. inermis. We explored the influence of geographic and environmental variables on genetic differ- Diversity and genetic structure of three Guadua species. entiation to test IBD and IBE in G. inermis. Linearized FST The average frequency of null alleles for each locus an-

Table 1. Genetic diversity statistics, genetic differentiation and inbreeding coefficient for three Guadua species for specimens from the states of Chiapas, Oaxaca, Puebla, and Veracruz, Mexico.

Species Code N N pop Ho HE GST GIS P

G. amplexifolia Ga 41 3 0.21 0.38 0.10 0.39 0.001 G. tuxtlensis Gt 16 2 0.33 0.38 0.04 0.06 0.09 G. inermis Gi 98 13 0.20 0.30 0.29 0.12 0.001

N sample size; N pop number of populations, Ho observed heterozygosity; HE expected heterozygosity; GST genetic differentiation; GIS inbreed- ing; P probability of GIS and Hardy-Weinberg equilibrium test.

547 Genetic structure of three Mexican bamboo species

Figure 3. A. STRUCTURE plot (K = 3 to 5) for the assignment of individuals. B. K groups identified by Evanno’s method. C. PCA for three species of Guadua based on five nuclear microsatellites.

alyzed was 0.26 for FJ444929, 0.08 for Aar12, 0.26 for orange) included individuals of G. amplexifolia, while the FJ476075, 0.30 for FJ444936, and 0.22 for FJ444930. The third and fourth groups (color: blue and purple respective- number of alleles for Guadua inermis, G. amplexifolia and ly) were composed of G. inermis plants. For these last two G. tuxtlensis was 28, 25, and 13 alleles, respectively. The groups, group 3 included mainly individuals from popula-

multilocus statistic of genetic diversity based on HO was tions Gi524, Gi537, Gi538, and Gi541, which are located 0.20 for G. inermis, 0.21 for G. amplexifolia and 0.33 for in the northern part of the sampling region and Gi548 lo- G. tuxtlensis (Table 1). Average levels of genetic diver- cated in the state of Puebla. Group 4 was composed of sity were lower for G. inermis, and most of the estimated plants from populations Gi466, Gi526, and Gi527, which statistics were higher for G. tuxtlensis (Appendix 2). The are in the central part of the study region. The southern rarefaction statistic for AR was 1.45 for G. inermis, 1.71 populations Gi544, Gi528, Gi529, Gi530 had plants as- for G. amplexifolia, and 1.71 for G. tuxtlensis. All popula- signed to groups 3 or 4 and some individuals had a mixed tions of G. amplexifolia (Ga519, Ga531, and Ga536) ex- ancestry. Population Gi474, located in Chiapas, had a hibited significant values of inbreeding, but this only oc- mixed ancestry from groups 2, 3, and 4; population Gi524 curred in three populations of G. inermis (Gi524, Gi528, also had mixed ancestry with group 2 that corresponds to and Gi529) and in one G. tuxtlensis population (Gt532) the G. amplexifolia cluster. The PCA for the three species (Appendix 2). had a pattern similar to that observed in STRUCTURE, According to Delta K, the STRUCTURE analysis where there is a distinction among the populations of the showed that the most likely number of clusters was K = three species (Figure 3). 4 (Figure 3). Group 1 (color: green) corresponded to in- The AMOVA indicated that 59.89 % of the genetic vari- dividuals of Guadua tuxtlensis; the second group (color: ance was explained among the three species, in contrast to

548 Pérez-Alquicira et al. / Botanical Sciences 99(3): 542-559. 2021

Table 2. Analysis of molecular variance for a) three species of Guadua and b) G. inermis. Source of variation df SS MS % CV

a) Among groups 2 107.24 53.62 59.89 Among populations within species 15 51.36 3.42 19.38 Within populations 104 51.72 0.49 20.71 Total 121 210.33 1.73 100 b) Among populations 12 111.93 9.32 71.74 Within populations 57 37.63 0.66 28.25 Total variation 69 149.57 2.16 100

19.39 % of the genetic variance among populations within found in population Gi537 located in the northern part species and 20.71 % of the variance within populations of the sampling region, and population Gi548 in Puebla,

(Table 2). The GST for G. amplexifolia, G. tuxtlensis and while the highest values of HO, HE, and NAE were detected G. inermis was 0.10, 0.04 and 0.29 respectively (Table 1). for Gi528 and Gi530 located in the central and southern The highest levels of genetic differentiation were detected part of the sampling region (Appendix 2). between one population of Guadua tuxtlensis (Gt480) According to Delta K for STRUCTURE, the most and the populations of G. inermis located in the northern likely number of clusters was K = 2. The first group (or- region of the sampling area in Veracruz (Gi537, Gi538, ange, Figure 4) was composed of five populations (Gi537, and Gi541; Appendix 3). The lowest levels of genetic Gi538, Gi541, Gi548) from the northern part of Veracruz differentiation were detected between Ga531 located in and one population (Gi474) from Chiapas. The second Oaxaca and Gi474 located in Chiapas; also Ga531 exhib- group included three populations (Gi466, Gi526, and ited low levels of genetic differentiation with populations Gi527) from the central part of the sampling region in of G. inermis located mainly in the southern part of the Veracruz, while five populations (Gi524, Gi544, Gi528, sampling region in Veracruz (Appendix 3). These results Gi529, and Gi530) included individuals with mixed an- agree with the PCA and STRUCTURE analyses, where G. cestry from the two groups, located mainly in southern amplexifolia and G. inermis had lower levels of genetic ranges (Figure 4). The PCA exhibited a pattern similar differentiation than G. tuxtlensis did. to the result of K = 3 from STRUCTURE, where some individuals are intermixed among populations. Specifi- Diversity, genetic structure, and connectivity of Guadua cally, one group included the most northern populations inermis. Average levels of genetic diversity of G. inermis Gi541, Gi474, Gi524, Gi548, Gi544, Gi537, and Gi538. were HE = 0.23 and Ho = 0.20 (Appendix 2). The lowest The second group included populations located mainly in levels of genetic diversity based on HO, HE, and NAE were southern ranges, Gi528, Gi528, Gi530 (purple group of

Table 3. Results from a multimodel inference on MLPE regression models in G. inermis. Log likelihood (LogLik), Akaike’s information criterion corrected for sample size (AICc), relative difference between the best model and each of the other models in the set (ΔAIC), model weight representing the probability that a model is the best in the set (wi), and the correlation coefficient rho ρ( ). Predictor logLik AICc ΔAIC wi Ρ Bio 5 29.90 -51.3 0 1 0.333* IBD 18.62 -28.7 22.57 0 0.109 Bio19 18.03 -27.5 23.73 0 0.096 Bio10 16.33 -25.1 26.12 0 0.041 Bio 3 16.28 -24.0 27.25 0 0.054 Bio 13 16.14 -23.7 27.54 0 0.050

*P = 0.0001.

549 Genetic structure of three Mexican bamboo species

K = 3) and the third group, those located in the central part tuxtlensis than in G. inermis or G. amplexifolia for most of of the sampling area, Gi527, and Gi526 (the same as the the statistics, while the lowest levels of genetic diversity pure blue group in K = 2 and 3) (Figure 4). The Gi474 pop- were detected in G. inermis. However, it was a population ulation (located in Oaxaca) included wide variation, with of G. inermis that exhibited the highest levels of genetic some individuals similar to other populations while others diversity, according to the Ho and HE statistics for the are divergent. The AMOVA results showed that 71.74 % of Gi530 population, which is in the southern region of Ve- the genetic variance was explained by differences among racruz. The PCA, STRUCTURE, and AMOVA indicated populations, while 28.25 % could be attributed­ to the ge- that each species is genetically different, supporting their netic variance within populations (Table 2). taxonomic designation. The PCA, however, did reveal that

G. inermis populations exhibited wide variation for FST G. amplexifolia and G. inermis are genetically closer to pairwise comparisons, ranging from 0 to 0.67. Genetic each other than to G. tuxtlensis. This result was supported differentiation at the species level was FST = 0.47 and GST by slightly lower FST pairwise comparisons between both = 0.44. The highest levels of genetic differentiation were species. According to Londoño (pers. comm.), G. inermis detected between populations located in Chiapas, those in and G. amplexifolia would be grouped in the same genetic the center of the sampling area in Veracruz (Gi466, Gi526, complex, while G. tuxtlensis would be grouped with other and Gi527), and the northernmost populations of Veracruz Guadua species. The inbreeding coefficient for most of (Gi537, Gi538, and Gi541). Overall, the lowest levels of the G. inermis populations we studied indicated an excess genetic variation (FST = 0) were found in geographically of heterozygotes, and the three populations of G. amplexi- close populations (Appendix 3). folia exhibited inbreeding. This result is partial, because We did not find a significant IBD pattern (r = 0.13, P = G. amplexifolia has a wide distribution (Mexico to Co- 0.2). Multimodal inference showed that the best perform- lombia) and we only analyzed three populations of this ing MPLE model included the maximum temperature of species in Mexico. the warmest month (Bio 5) as this variable had the maxi- mum model probability (wi = 1) and the lowest AICc score Genetic diversity and connectivity in G. inermis. Multiple (-51.3) relative to the other models. This environmental factors determine the genetic diversity of species, including variable explained 33 % of the genetic variation in popula- flowering synchrony, geographic range, and mating system tion genetic distances (Table 3). None of the other biocli- (Hamrick & Godt 1996, Nybom 2004). Genetic diversity mate variables explained gene flow inG. inermis. in bamboos exhibits a wide range of variation. Guadua inermis harbors, on average, lower levels of genetic di- Discussion versity (Ho = 0.20 and He = 0.23) compared to six other

species of bamboos using SSR (Ho = 0.55 and HE = 0.46) Our main results confirm the taxonomic designation of the (Posso 2011, Attigala et al. 2017, Jiang et al. 2017, Yang three Guadua species as they exhibited evident genetic dif- et al. 2018, Meena et al. 2019, Huang et al. 2020). Among ferentiation (Londoño & Ruiz-Sanchez 2014). This was the highest levels of genetic diversity detected were those clear from the results of STRUCTURE and PCA that co- of G. angustifolia (HE = 0.56 including 30 locations and 9 herently separated the three species. Pairwise genetic dif- SSR markers in the Colombian Eje Cafetalero; Muñoz et al. ferentiation among species was also high. The average lev- 2010, Posso 2011). els of genetic diversity were lower for G. inermis, and the G. angustifolia has a wide distribution range with mas- values for most of the statistics were larger for G. tuxtlen- sive and sporadic flowering; factors that could explain the sis. We found a strong genetic structure in G. inermis. We high levels of genetic diversity in this species. In contrast, found a significant association of genetic distances with the even though the bamboo sinicus Chia & maximum temperature of the warmest month, but not with J.L. Sun, flowers sporadically and has a narrow distribu- geographic distance. Moreover, we did not find any evi- tion, it has higher levels of genetic diversity than G. iner- dence supporting the hybrid origin of G. tuxtlensis (which mis does (HE = 0.54, HO = 0.48, including 8 SSR markers was thought to have G. inermis as one of its parental spe- and 18 populations; Yang et al. 2018). Chen et al. (2017) cies), as suggested by Londoño & Ruiz-Sánchez (2014). found high levels of outcrossing rates in this species and another study estimated low levels of biparental inbreed- Genetic diversity and structure among Guadua species. ing in D. sinicus (Xie et al. 2019). This explains the high Higher levels of genetic diversity were detected in Guadua values of diversity detected in this species. Other bamboos

550 Pérez-Alquicira et al. / Botanical Sciences 99(3): 542-559. 2021

Figure 4. A. STRUCTURE bar plots (K = 2 and 3) for the assignment of individuals. B. K groups identified by Evanno’s method. C. PCA for 13 popula- tions of Guadua inermis based on five microsatellites. also exhibit high levels of genetic diversity and excess monocot outcrossing species, their levels of genetic di- heterozygosity even when flowering is sporadic (Phyl- versity are higher (HE = 0.15; Hamrick & Godt 1996). lostachys edulis - Jiang et al. 2017, Fargesia spathacea Polyploidy has been one of the factors that contribute to complex - Huang et al. 2020). The authors suggested that maintaining high levels of genetic diversity in bamboos. even the sporadic flowering events can potentially con- Theoretical and experimental data have shown a reduction tribute to the population diversity of these species. of inbreeding in polyploid species as a consequence of a Moreover, the flowering time interval within the spo- higher number of alleles per locus (Moody et al. 1993, radic phenology should vary among species, which po- Soltis & Soltis 2000, Baduel et al. 2018). tentially contributes to the variation in genetic diversity The higher genetic diversity and mixed ancestry of among bamboo species. In the case of G. inermis, dur- southern populations of Guadua inermis suggest that this ing sample collection, we detected only six individuals region is the ancestral distribution. This result agrees with flowering in six different populations. This highlights just the southern origin of tropical bamboos (Clark 1997, Ruiz- how rare flowering is. However, it is necessary to further Sanchez 2011). Northern range expansion could occur as a analyze flowering phenology to understand its role in the consequence of warming weather in the Late Pleistocene structure of population genetics. and Holocene. Other species inhabiting this region also G. inermis has lower levels of genetic diversity than migrated toward northern ranges (Ruiz-Sanchez & Orne- other bamboo species. When it is compared to 79 other las 2014, Ornelas et al. 2016, 2019).

551 Genetic structure of three Mexican bamboo species

The AMOVA indicated high levels of genetic differ- cally close populations are genetically differentiated, par- entiation among populations. Further, we detected large ticularly in the center and northern ranges. genetic differences between northern and central popula- Our study is the first to evaluate levels of genetic di- tions, even when they are in proximity. Further research, versity in Mexican endemic woody bamboos. We detected including the use of more genetic markers, is needed to that the genetic diversity of G. inermis is on average lower test this result. High genetic structure has been also de- than that reported in six other bamboo studies using SSR, tected in the tallest bamboo Dendrocalamus giganteus us- and higher than that of other outcrossing monocots. Even ing ISSR (7 primers, 7 populations, GST = 0.84; Tian et with its small population size, patchy distribution and al. 2012). The authors indicated that intensive selection of its sporadic flowering, G. inermis maintains its genetic genotypes and their introduction into natural populations diversity. Its sporadic flowering and habitat fragmenta- is one of the reasons for the high levels of genetic differen- tion could reduce genetic connectivity for this species. tiation. Moreover, the long vegetative period and sporadic We detected a strong genetic structure between northern flowering are also important factors that influence the high and central populations. Even when populations are geo- degree of genetic structure and low levels of genetic diver- graphically close, their genetic differences are as high as sity (HE = 0.04 and HO = 0.06; Tian et al. 2012). those in the comparisons of populations among species. Guadua inermis had higher levels of genetic differ- Our study also confirms the taxonomic identity of three entiation than G. angustifolia (FST = 0.098; Posso 2011), endemic bamboo species.

D. sinicus (GST = 0.23; Yang et al. 2018), or D. membra- naceus did (21.05 % of the variation was among popula- Acknowledgments tions; Yang et al. 2012). Even though D. sinicus has spo- radic flowering, its genetic structure was not greater than We extend our thanks to Ariel Muñoz Sánchez and Pilar that of G. inermis. It would be important to increase the Zamora-Tavares, for technical assistance in the labora- number of genetic markers in future studies to increase tory and to Diego Angulo, Bill Wysocki, Etelvina Gánda- accuracy when quantifying the levels of genetic diversity ra Zamorano, Andrés Ortíz Rodríguez, Ismael Guzman and structure of these species. Valdivieso, Regina Cuevas and Víctor Piñeros for their Anthropogenic habitat loss and fragmentation con­ help in the field collections. Our project was supported by tribute to the reduction in population size and isolation of the UDG and CONACyT-Labo­ratorio Nacional de Iden- populations (Reed & Frankham 2003, Lowe et al. 2005, tificación y Caracterización Vegetal (LaniVeg) (Grant Chávez-Pesqueira et al. 2014, Schlaepfer et al. 2018), num. 293833) for laboratory work; and CONACyT to which endemic species are more susceptible. Addi- (Grant num. 215514) for field trips. Bianca Delfosse edited tionally, the loss of genetic diversity could have a nega- the English version of this manuscript. tive impact on fitness, viability of individuals, and their ability to respond to environmental challenges (Reed & Literature cited Frankham 2003). Habitat fragmentation, the sporadic flowering and the projected future reduction in the dis- Abreu AG, Grombone-Guaratini MT, Monteiro M, Pin- tribution of G. inermis (Ruiz-Sanchez et al. 2018) all heiro JB, Tombolato AFC, Zucchi MI. 2011. Develop- threaten the prevalence of this species. Genetic diversity ment of microsatellite markers for Aulonemia aristu- is an important aspect in the conservation and utilization lata (Poaceae) and cross-amplification in other bamboo of resources. Our study contributes to this knowledge species. American Journal of Botany 98: e90-2. DOI: and can be helpful when drawing up conservation man- https://doi.org/10.3732/ajb.1000511 agement plans. Aguilera López S. 2020. Estructura genética de Guadua Precipitation and temperature are important factors in- inermis (Poaceae:Bambusoideae), especie endémica de fluencing plant growth, development, survival, reproduc- la vertiente del golfo de México. BSc Thesis. Universi- tion, and other ecological aspects (Manel et al. 2012). Our dad de Guadalajara,. results show that the environment influenced the genetic Attigala L, Gallaher T, Nason J, Clark LG. 2017. Genetic structure of G. inermis. Specifically, the maximum tem- diversity and population structure of the threatened perature of the warmest month was significantly correlated temperate woody bamboo Kuruna debilis (Poaceae: with genetic distances, while geographic distance had no Bambusoideae: Arundinarieae) from Sri Lanka based effect. This result was partially expected since geographi- on microsatellite analysis. Journal of the National Sci-

552 Pérez-Alquicira et al. / Botanical Sciences 99(3): 542-559. 2021

ence Foundation of Sri Lanka 45: 53-5. DOI: https:// Cortés-Rodríguez GR. 2000. Los bambúes nativos de doi.org/10.4038/jnsfsr.v45i1.8038 México. CONABIO. Biodiversitas 30: 12-15. Baduel P, Bray S, Vallejo-Marin M, Kolár F, Yant L. 2018. De Silva HN, Hall AJ, Rikkerink E, McNeilage MA, The “Polyploid Hop”: shifting challenges and oppor- Fraser LG. 2005. Estimation of allele frequencies in tunities over the evolutionary lifespan of genome du- polyploids under certain patterns of inheritance. He- plications. Frontiers in Ecology and Evolution 6: 1-19. redity 95: 327-334. DOI: https://doi.org/10.1038/ DOI: https://doi.org/10.3389/fevo.2018.00117 sj.hdy.6800728 Banik RL. 1998. Reproductive biology and flowering Domínguez CA, Dirzo R. 1995. Rainfall and flowering populations with diversities in muli bamboo, Melo- synchrony in a tropical shrub: Variable selection on the canna baccifera (Roxb.) Kurz. Bangladesh Journal of flowering time of Erythroxylum havanense. Evolution- Forest Science 27: 1-15. ary Ecology 9: 204-216. DOI: https://doi.org/10.1007/ Bartoń K. 2018. MuMIn: multi-model inference. R pack- BF01237757 age version 1.4. Doyle JJ, Doyle JJ. 1987. A rapid DNA isolation proce- Bhattacharya S, Ghosh JS, Das M, Pal A. 2009. Morpho- dure for small quantities of fresh leaf tissue. Phyto- logical and molecular characterization of Thamno- chemical Bulletin 19: 11-15. calamus spathiflorus subsp. spathiflorus at population Earl D, vonHoldt BM. 2012. STRUCTURE HARVEST- level. Plant Systematic and Evolution 282: 13-20. DOI: ER: A website and program for visualizing STRUC- https://doi.org/10.1007/s00606-008-0092-1 TURE output and implementing the Evanno method. Bogdziewicz M, Pesendorfer M, Crone EE, Pérez-Izqui- Conservation Genetics 4: 359-361. DOI: https://doi. erdo C, Bonal R. 2020. Flowering synchrony drives re- org/10.1007/s12686-011-9548-7 productive success in a wind-pollinated tree. Ecology Esselink GD, Nybom H, Vosman B. 2004. Assignment of Letters 23: 1820-1826. DOI: https://doi.org/10.1111/ allelic configuration in polyploids using the MAC-PR ele.13609 (microsatellite DNA allele counting-peak ratios) meth- Chávez-Pesqueira M, Suárez Montes P, Castillo G, od. Theoretical and Applied Genetics 109: 402-408. Núñez-Farfán J. 2014. Habitat fragmentation threatens DOI: https://doi.org/10.1007/s00122-004-1645-5 wild populations of Carica papaya (Caricaceae) in a Evanno G, Regnaut S, Goudet J. 2005. Detecting the lowland rainforest. American Journal of Botany 101: number of clusters of individuals using the software 1092-1101. DOI: https://doi.org/10.3732/ajb.1400051 structure: A simulation study. Molecular Ecology Chen LN, Cui YZ, Wong KM, Li DZ, Yang HQ. 2017. 14: 2611-2620. DOI: https://doi.org/10.1111/j.1365- Breeding system and pollination of two closely related 294X.2005.02553.x bamboo species. AoB Plants 9: plx021. DOI: https:// Fitter AH, Fitter RSR. 2002. Rapid changes in flowering doi.org/10.1093/aobpla/plx021 time in British plants. Science 296: 1689-1691. DOI: Clark LG. 1997. Bamboos: the centerpiece of the grass https://doi.org/10.1126/science.1071617 family. In: Chapman GP. ed. The Bamboos. London: Franklin DC. 2004. Synchrony and asynchrony: obser- Academic Press, pp. 237-248. ISBN: 9780674428683 vations and hypothesis for the flowering wave in a Clark LG, Londoño X, Ruiz-Sanchez E. 2015. Bamboo long-lived semelparous bamboo. Journal of Bioge- taxonomy and habitat. In: Liese W, Köhl M, eds. Bam- ography 31: 773-786. DOI: https://doi.org/10.1111/ boo: The plant and its uses. Switzerland: Springer In- j.1365-2699.2003.01057.x ternational Publishing, pp. 1-30. DOI: https://dx.doi. Gadgil M, Prasad SN. 1984. Ecological determinants of life org/10.1007/978-3-319-14133-6_1 history evolution of two Indian bamboo species. Biotro- Clark LV, Jasieniuk M. 2011. POLYSAT: An R pack- pica 16: 161-172. DOI: https://doi.org/10.2307/2388050 age for polyploidy microsatellite analysis. Molecu- Guerreiro C. 2014. Flowering cycles of woody bamboos lar Ecology Resources 11: 562-566. DOI: https://doi. native to southern South America. Journal of Plant org/10.1111/j.1755-0998.2011.02985.x Research 127: 307-313. DOI: https://doi.org/10.1007/ Clarke RT, Rothery P, Raybould AF. 2002. Confidence lim- s10265-013-0593-z its for regression relationships between distance matri- Guerreiro C, Peichoto M, Vega AS. 2020. Flowering of ces: estimating gene flow with distance.Journal of Ag- Guadua trinii (Poaceae, Bambusoideae, Bambuse- ricultural, Biological, and Environmental Statistics 7: ae). Rodriguésia 71: e03352018. DOI: https://doi. 361-372. DOI: https://doi.org/10.1198/108571102320 org/10.1590/2175-7860202071033

553 Genetic structure of three Mexican bamboo species

Guo ZH, Ma PF, Yang GQ, Hu JY, Liu YL, Xia EH, der assortive mating. Proceedings of the Royal Society Zhao L, Sun GL, Xu YX, Zhao Yj, Zhang. YC, Zhang of London. Series B: Biological Sciences 267: 1649- XM, Zhou MY, Guo Y, Guo C, Liu JX, Ye XY, Chen 1655. DOI: https://doi.org/10.1098/rspb.2000.1191 YM, Yang Y, Han B, Lin CS, Lu Y, Li DZ. 2019. Ge- Kopelman NM, Mayzel J, Jakobsson M, Rosenberg nome sequences provide insights into the reticulate NA, Mayrose I. 2015. CLUMPAK: a program for origin and unique traits of woody Bamboos. Molecular identifying clustering modes and packaging popula- Plant 12: 1353-1365. DOI: https://doi.org/10.1016/j. tion structure inferences across K. Molecular Ecol- molp.2019.05.009 ogy Resources 15: 1179-1191. DOI: https://doi. Hamrick JL, Godt MJW. 1996. Effects of life history traits org/10.1111/1755-0998.12387 on genetic diversity in plant species. Philosophical Liebsch D, Reginato M. 2009. Florescimento e frutifica- Transactions of the Royal Society of London. Series B: ção de Merostachys skvortzovii Sendulsky (taquara- Biological Sciences 351: 1291-1298. DOI: https://doi. lixa) no estado do Paraná. Iheringia. Série Botânica org/10.1098/rstb.1996.0112 64: 53-56. Hardy OJ, Vekemans X. 2002. SPAGeDi: a versatile com- Londoño X. 2001. Guadua Kunth. In: Wanderley MGL, puter program to analyse spatial genetic structure at Shepherd G, Giulietti AM, eds. Poaceae, Flora fanero- the individual or population levels. Molecular Ecology gâmica do estado de São Paulo. São Paulo: Hucitec, pp Notes 2: 618-620. DOI: https://doi.org/10.1046/j.1471- 38-39. ISBN: 85-7523-051-4. 8286.2002.00305.x Londono X. 2002. Distribucion, morfología, taxonomía, Hijmans RJ. 2019. Geosphere: Spherical Trigonometry. R anatomía, silvicultura y usos de los bambúes del Nuevo package version 1.5-10. Mundo. Universidad Nacional de Colombia, Santa Fe Huang L, Xing XC, Li WW, Zhou Y, Zhang YQ, Xue de Bogota. http://www.maderinsa.com/guadua/taller. C, Ren Y, Kang JQ. 2020. Population genetic struc- html (accessed November 3, 2020). ture of the giant panda staple food bamboo (Fargesia Londoño X. 2011. El Bambú en Colombia. Biotecnología spathacea complex) and its taxonomic implications. Vegetal 11: 143-154. Journal of Systematics and Evolution DOI: https://doi. Londoño X, Peterson PM. 1991. Guadua sarcocarpa (Po- org/10.1111/jse.12594 aceae: Bambuseae), a new species of Amazonian bam- Janzen DH. 1976. Why bamboos wait so long to flow- boo with fleshy fruits. Systematic Botany 16: 630-638. er. Annual Review of Ecology and Systematics 7: DOI: https://doi.org/10.2307/2418866 347-391. DOI: https://doi.org/10.1146/annurev. Londoño X, Ruiz-Sanchez E. 2014. Guadua tuxtlensis es.07.110176.002023 (Poaceae: Bambusoideae: Bambuseae: Guaduinae): Jiang W, Bai T, Dai H, Wei Q, Zhang W, Ding Y. 2017. Una nueva especie inadvertida de la región de Los Microsatellite markers revealed moderate genetic di- Tuxtlas, Veracruz, México. Botanical Sciences 92: versity and population differentiation of moso bamboo 481-488. DOI: https://doi.org/10.17129/botsci.76 (Phyllostachys edulis)-a primarily asexual reproduc- Lowe AJ, Boshier D, Ward M, Bacles CFE, Navarro C. tion species in China. Tree Genetics & Genomes 13: 2005. Genetic resource impacts of habitat loss and deg- 130. DOI: https://doi.org/10.1007/s11295-017-1212-2 radation; reconciling empirical evidence and predicted Jombart T, Devillard S, Balloux F. 2010. Discriminant theory for neotropical trees. Heredity 95: 255-273. analysis of principal components: a new method for DOI: https://doi.org/10.1038/sj.hdy.6800725 the analysis of genetically structured populations. BMC Manel S, Gugerli F, Thuiller W, Alvarez N, Legendre P, Genetics 11: 94. DOI: https://doi.org/10.1186/1471- Holderegger R, Gielly L, Taberlet P, IntraBioDiv Con- 2156-11-94 sortium. 2012. Broad-scale adaptive genetic variation Judziewicz EJ, Clark LG, Londoño X, Stern MJ. 1999. in alpine plants is driven by temperature and precipita- American Bamboos. Washington, DC: Smithsonian In- tion. Molecular Ecology 21: 3729-3738. DOI: https:// stitution Press. ISBN: 1560985690. doi.org/10.1111/j.1365-294X.2012.05656.x Kamvar ZN, Tabima JF, Grünwald NJ. 2014. Poppr: an R Medina JC, Medina DM. 1965. Observacoes citológicas package for genetic analysis of populations with clon- em Guadua superba Huber. Bragantia. 12: 63-68. al, partially clonal, and/or sexual reproduction. PeerJ Meena RK, Bhandhari MS, Barhwal S, Ginwal HS. 2019. 2: e281. DOI: https://doi.org/10.7717/peerj.281 Genetic diversity and structure of Dendrocalamus ham- Kirkpatrick M. 2000. Reinforcement and divergence un- iltonii natural metapopulation: a commercially impor-

554 Pérez-Alquicira et al. / Botanical Sciences 99(3): 542-559. 2021

tant bamboo species of northeast Himalayas. 3 Biotech Reed DH, Frankham R. 2003. Correlation between fitness 9: 60. DOI: https://doi.org/10.1007/s13205-019-1591-1 and genetic diversity. Conservation Biology 17: 230- Meirmans PG. 2020. GENODIVE version 3.0: Easy-to-use 237. software for the analysis of genetic data of diploids and Reisch C, Poschlod P. 2009. Land use affects flowering polyploids. Molecular Ecology Resources 20: 1126- time: seasonal and genetic differentiation in the grass- 1131. DOI: https://doi.org/10.1111/1755-0998.13145 land plant Scabiosa columbaria. Evolutionary Ecol- Moody ME, Mueller LD, Soltis DE. 1993. Genetic varia- ogy 23: 753-764. DOI: https://doi.org/10.1007/s10682- tion and random drift in autotetraploid populations. 008-9270-4 Genetics 134: 64-657. Rodríguez-Pérez J, Traveset A. 2016. Effects of flowering Muñoz JE, Londoño X, Rugeles P, Posso AM, Vallejo FA. phenology and synchrony on the reproductive success 2010. Diversidad y estructura genética de Guadua an- of a long-flowering shrub.AOB Plants 8: plw007. DOI: gustifolia en la ecorregión cafetera colombiana. Recur- https://doi.org/10.1093/aobpla/plw007 sos Naturales y Ambiente 61: 45-52 Rosenberg NA. 2003. Distruct: a program for the graphi- Naimi B. 2015. usdm: Uncertainty analysis for species cal display of population structure. Molecular Ecology distribution models. R package version, 1.1.18. Notes 4: 137-138. DOI: https://doi.org/10.1046/j.1471- Nybom H. 2004. Comparison of different nuclear DNA 8286.2003.00566.x markers for estimating intraspecific genetic diversity Rousset F. 1997. Genetic differentiation and estimation of in plants. Molecular Ecology 13: 1143-1155. DOI: gene flow from F-statistics under isolation by distance. https://doi.org/10.1111/j.1365-294X.2004.02141.x Genetics 145: 1219-28. Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin Ruiz-Sanchez E. 2011. Biogeography and divergence time PR, O’Hara RB, Simpson GL, Solymos P, Stevens estimates of woody bamboos: insights in the evolution MHH, Wagner H, Oksanen MJ. 2013. Package vegan. of Neotropical bamboos. Botanical Sciences 88: 67-75. Community Ecology Package, version 2: pp.1-295. DOI: https://doi.org/10.17129/botsci.312 Ornelas JF, Gándara E, Vásquez-Aguilar AA, Ramírez- Ruiz-Sanchez E, Clark LG, Londoño X, Mejía-Saulés Barahona S, Ortiz-Rodriguez AE, González C, Me- T, Cortés G. 2015. Morphological keys to the genera jía-Saules T, Ruiz-Sanchez E. 2016. A mistletoe tale: and species of bamboos (Poaceae: Bambusoideae) postglacial invasion of Psittacanthus schiedeanus (Lo- of Mexico. Phytotaxa 236: 1-24. DOI: https://doi. ranthaceae) to Mesoamerican cloud forests revealed org/10.11646/phytotaxa.236.1.1 by molecular data and species distribution modeling. Ruiz-Sanchez E, Mendoza-Gonzalez G, Rojas-Soto O. BMC Evolutionary Biology 16: 78. DOI: https://doi. 2018. Mexican priority bamboo species under scenar- org/10.1186/s12862-016-0648-6 ios of climate change. Botanical Sciences 96: 11-23. Ornelas JF, Ortiz-Rodriguez AE, Ruiz-Sanchez E, Sosa DOI: https://doi.org/10.17129/botsci.1206 V, Pérez-Farrera MÁ. 2019. Ups and downs: genetic Ruiz-Sanchez E, Munguía-Lino G, Vargas-Amado G, differentiation among populations of the Podocarpus Rodríguez A. 2020. Diversity, endemism and conser- (Podocarpaceae) species in Mesoamerica. Molecular vation status of native Mexican woody bamboos (Poa- Phylogenetics and Evolution 138: 17-30. DOI: https:// ceae: Bambusoideae: Bambuseae). Botanical Journal doi.org/10.1016/j.ympev.2019.05.025 of the Linnean Society 192: 281-295. DOI: https://doi. Pérez-Galindo P, Carlos Andrés C, Ginzález G, Iván An- org/10.1093/botlinnean/boz062 drés I, Cárdenas C, Heiber H. 2009. Cloning and isola- Ruiz-Sanchez E, Ornelas J F. 2014. Phylogeography of tion of tetra nucleotide microsatellite clones from Gua- Liquidambar styraciflua (Altingiaceae) in Mesoamer- dua angustifolia (Poaceae: Bambusoideae). Molecular ica: survivors of a Neogene widespread temperate for- Ecology Resources 9: 1375-1379. est (or cloud forest) in North America? Ecology and Pritchard JK, Stephens M, Donnelly P. 2000. Inference of Evolution 4: 311-328. DOI: https://doi.org/10.1002/ population structure from multilocus genotype data. ece3.938 Genetics 155: 945-959. Ruiz-Sanchez E, Tyrrel CD, Londoño X, Oliveira RP, Posso T. 2011. Diversidad Genética y Estructura Pobla- Clark LG. 2021. Diversity, distribution, and classifica- cional de Guadua angustifolia Kunth en el eje cafetero tion of Neotropical woody bamboos (Poaceae: Bam- Colombiano. MSc Thesis, Universidad Nacional de busoideae) in the 21st Century. Botanical Sciences 99: Colombia. 198-228. DOI: https://doi.org/10.17129/botsci.2722

555 Genetic structure of three Mexican bamboo species

Schlaepfer DR, Braschler B, Rusterholz HP, Baur B. bamboo (Poaceae: Bambusoideae: Bambuseae) hidden 2018. Genetic effects of anthropogenic habitat frag- by convergent character evolution. Taxon 67: 916-930. mentation on remnant animal and plant populations: a DOI: https://doi.org/10.12705/675.5 meta-analysis. Ecosphere 9: e02488. DOI: https://doi. Vega AS, Hernández JC. 2008. La floración de Guadua org/10.1002/ecs2.2488 chacoensis (Poaceae, Bambusoideae, Bambuseae). Re- Slatkin M. 1993. Isolation by distance in equilibrium vista de la Facultad de Agronomía 28: 107-110. and non-equilibrium populations. Evolution 47: 264- Wang IJ, Bradburg GS. 2014. Isolation by environment. 279. DOI: https://doi.org/10.1111/j.1558-5646.1993. Molecular Ecology 23: 5649-5662. DOI: https://doi. tb01215.x org/10.1111/mec.12938 Soltis PS, Soltis DE. 2000. The role of genetic and ge- Xie N, Chen L, Dong Y, Yang H. 2019. Mixed mating nomic attributes in the success of polyploids. Proceed- system and variable mating patterns in tropical woody ings of the National Academy of Sciences of the United bamboos. BMC Plant Biology 19: 418. DOI: https:// States of America 97: 7051-7057. DOI: https://doi. doi.org/10.1186/s12870-019-2024-3 org/10.1073/pnas.97.13.7051 Yang JB, Dong YR, Wong KM, Gu ZJ, Yang HQ, Li DZ. Szpiech ZA, Jakobsson M, Rosenberg NA. 2008. ADZE: A 2018. Genetic structure and differentiation in Dendro- rarefaction approach for counting alleles private to com- calamus sinicus (Poaceae: Bambusoideae) populations binations of populations. Bioinformatics 24: 2498-2504. provide insight into evolutionary history and speciation DOI: https://doi.org/10.1093/bioinformatics/btn478 of woody bamboo. Scientific Reports 8: 16933. DOI: Tian B, Yang HQ, Wong KM, Liu AZ, Ruan ZY. 2012. https://doi.org/10.1038/s41598-018-35269-8 ISSR analysis shows low genetic diversity versus high Yang HQ, An MY, Gu ZJ, Tian B. 2012. Genetic Diversity genetic differentiation for giant bamboo, Dendrocala- and Differentiation of Dendrocalamus membranaceus mus giganteus (Poaceae: Bambusoideae), in China (Poaceae: Bambusoideae), a Declining Bamboo Spe- populations. Genetic Resources and Crop Evolution cies in Yunnan, China, as Based on Inter-Simple Se- 59: 901-908. DOI: https://doi.org/10.1007/s10722- quence Repeat (ISSR) Analysis. International Journal 011-9732-3 of Molecular Sciences 13: 4446-4457. DOI: https://doi. Troup RS. 1921. The silviculture of Indian trees. Volume org/10.3390/ijms13044446 1. Oxford, UK: Oxford University Press. Zheng X, Lin S, Fu H, Wan Y, Ding Y. 2020. The bam- Tyrrell CD, Londoño X, Prieto RO, Attigala L, McDonald boo flowering cycle sheds light on flowering diversity. K, Clark LG. 2018. Molecular phylogeny and cryptic Frontiers in Plant Science 11: 381. DOI: https://doi. morphology reveal a new genus of West Indian woody org/10.3389/fpls.2020.00381

Associate editor: Monserrat Vázquez Sánchez Authors’ contributions: JPA carried out lab work, data analyses and wrote the manuscript. SAL carried out lab work and reviewed the manuscript. YR, participated in data analyses, wrote and reviewed the manuscript. ERS conceived and designed the study, collected the samples, and wrote and reviewed the manuscript.

556 Pérez-Alquicira et al. / Botanical Sciences 99(3): 542-559. 2021

Appendix 1. Population code (Code), location, longitude and latitude, elevation and voucher information for three Guadua species from eastern Mexico.

Code Location Longitude Latitude Elevation Voucher N W (m asl) Ga519 Veracruz -96.387222 19.492222 23 E. Ruiz-Sanchez et al. 519 Ga536 Veracruz -96.390227 19.536512 24 E. Ruiz-Sanchez & I. Guzmán 536 Ga531 Oaxaca -94.993222 16.738083 289 E. Ruiz-Sanchez & A. Ortíz 531 Gi466 Veracruz -96.156139 18.572472 6 E. Ruiz-Sanchez & W. Wysocki 466 Gt480 Veracruz -95.096293 17.575046 54 E. Ruiz-Sanchez & A. Ortíz 480 Gt532 Veracruz -94.826361 17.957167 32 E. Ruiz-Sanchez & A. Ortíz 532 Gi474 Chiapas -92.00225 15.892222 653 E. Ruiz-Sanchez & E. Gándara 474 Gi524 Veracruz -96.308028 18.668417 13 E. Ruiz-Sanchez & A. Ortíz 524 Gi526 Veracruz -96.100111 18.532306 6 E. Ruiz-Sanchez & A. Ortíz 526 Gi527 Veracruz -95.978639 18.434972 5 E. Ruiz-Sanchez & A. Ortíz 527 Gi528 Veracruz -95.709753 18.267389 3 E. Ruiz-Sanchez & A. Ortíz 528 Gi529 Veracruz -95.587222 18.188528 8 E. Ruiz-Sanchez & A. Ortíz 529 Gi530 Veracruz -94.973028 17.867306 7 E. Ruiz-Sanchez & A. Ortíz 530 Gi537 Veracruz -96.412369 19.671034 45 E. Ruiz-Sanchez & I. Valdivieso 537 Gi538 Veracruz -96.475312 19.867952 15 E. Ruiz-Sanchez & I. Valdivieso 538 Gi541 Veracruz -96.912586 19.36973 934 E. Ruiz-Sanchez & V. Piñeros 541 Gi544 Veracruz -94.785278 17.934722 21 E. Ruiz-Sanchez & A. Ortíz 544 Gi548 Puebla -98.373861 18.639056 1348 E. Ruiz-Sanchez & D. Angulo 548

557 Genetic structure of three Mexican bamboo species

Appendix 2. Genetic diversity parameters and inbreeding coefficient per population for threeGuadua (G. amplexifolia, G. inermis, and G. tuxtlensis) species from the states of Chiapas, Oaxaca, Puebla and Veracruz, Mexico.

Pop Location N NA NAe AR AR(ss4) HO HE GIS P G. inermis Gi466 VER 10 1.4 1.23 1.27 1.21 0.14 0.12 -0.20 0.02 Gi474 CHIS 6 2.4 1.82 1.97 1.67 0.21 0.31 0.33 0.23 Gi524 VER 5 2 1.38 1.7 1.47 0.16 0.25 0.36 0.00 Gi526 VER 4 1.8 1.51 1.67 1.50 0.26 0.27 0.02 0.52 Gi527 VER 4 1.6 1.29 1.48 1.30 0.22 0.18 -0.21 0.01 Gi528 VER 9 2.4 2.17 2.16 1.86 0.28 0.43 0.34 0 Gi529 VER 10 2.6 1.82 1.97 1.66 0.29 0.32 0.11 0.1 Gi530 VER 7 3 1.94 2.27 1.83 0.42 0.41 -0.03 0.28 Gi537 VER 7 1.2 1.13 1.19 1.14 0.10 0.07 -0.28 0.01 Gi538 VER 9 1.6 1.15 1.27 1.18 0.12 0.09 -0.23 0.00 Gi541 VER 16 2 1.16 1.31 1.20 0.13 0.10 -0.23 0 Gi544 VER 9 2.6 1.87 2.04 1.74 0.25 0.38 0.32 0.05 Gi548 PUE 2 1.2 1.14 1.2 1.15 0.1 0.08 -0.16 0.56 Average 1.98 1.50 1.6 1.45 0.20 0.23

G. amplexifolia Ga519 VER 15 3.4 1.74 2.07 1.68 0.23 0.33 0.31 0 Ga531 OAX 15 3.6 2.06 2.45 1.90 0.26 0.41 0.38 0 Ga536 VER 11 3.2 1.70 1.88 1.56 0.12 0.26 0.55 0 Average 3.4 1.83 2.1 1.71 0.20 0.33

G. tuxtlensis

Gt480 VER 8 2.2 1.68 1.81 1.62 0.38 0.33 -0.15 0.03 Gt532 VER 8 2.4 2.25 2.18 1.80 0.29 0.38 0.25 0.00 Average 2.3 1.96 1.99 1.71 0.33 0.35

N = sample size; NA = number of alleles; NAe = effective number of alleles; AR = allelic richness; Ho = observed heterozygosity;

HE = expected heterozygosity; GIS = inbreeding; P = probability of GIS and Hardy-Weinberg equilibrium test. VER = Veracruz, OAX = Oaxaca, CHIS = Chiapas, PUE = Puebla.

558 Pérez-Alquicira et al. / Botanical Sciences 99(3): 542-559. 2021

Appendix 3. Pairwise genetic differentiation parameter of three species of Guadua (G. amplexifolia, G. inermis, and G. tuxtlensis) for specimens from the states of Chiapas, Oaxaca, Puebla and Veracruz, Mexico. 466 474 524 526 527 528 529 530 537 538 541 544 548 519 531 536 480

474 0.48

524 0.54 0.20

526 0.30 0.38 0.41

527 0.15 0.31 0.40 0.15

528 0.28 0.16 0.21 0.12 0.16

529 0.37 0.07 0.23 0.34 0.27 0.13

530 0.17 0.14 0.22 0.08 0.09 0.02 0.10

537 0.67 0.45 0.04 0.64 0.63 0.37 0.38 0.38

538 0.66 0.47 0.05 0.63 0.62 0.38 0.39 0.39 0

541 0.65 0.50 0.08 0.64 0.62 0.43 0.43 0.43 0 0

544 0.39 0.20 0.11 0.15 0.24 0 0.21 0.10 0.26 0.24 0.28

548 0.65 0.27 0 0.53 0.58 0.20 0.28 0.27 0 0 0 0.05

519 0.61 0.43 0.43 0.54 0.52 0.44 0.42 0.43 0.59 0.60 0.64 0.48 0.52

531 0.51 0.32 0.32 0.45 0.41 0.35 0.34 0.33 0.50 0.51 0.56 0.40 0.41 0.02

536 0.68 0.46 0.44 0.59 0.58 0.50 0.51 0.50 0.63 0.64 0.67 0.51 0.54 0.22 0.12

480 0.68 0.58 0.60 0.47 0.60 0.46 0.55 0.44 0.73 0.73 0.76 0.53 0.66 0.59 0.52 0.62

532 0.65 0.45 0.49 0.44 0.55 0.39 0.47 0.38 0.64 0.64 0.68 0.44 0.53 0.53 0.47 0.57 0.06

559